A Novel Food Freshness Assay Based on ATP Depletion
Part 1: Automation and Sample Prep Optimization
August 08, 2014
Authors: Wendy Goodrich, Applications Department, BioTek Instruments, Inc., Winooski, VT, USA; Larissa Balakireva, NovoCIB, Lyon, France
Based on detection of products formed during the post-mortem degradation pathway of ATP (adenosine triphosphate), a novel mix and read assay has been developed for determining freshness in animal tissue. Inosine monophosphate (IMP), inosine (Ino), and hypoxanthine (Hx) are extracted from sample tissue and then converted to NADH using proprietary enzyme reagents. Expressed as a percent ratio value, calculation of relative content of the nucleotides is enabled by detection of NADH at 340 nm. In this first study, automated and non-automated performance parameters of the assay that could contribute to workflow or throughput enhancements in freshness testing were comparatively assessed, including instrument parameters and sample preparation options. Data including precision, accuracy, linearity, inter- and intra-assay repeatability, and sensitivity where applicable are presented. As reported in a companion Application Note, assay challenge testing was undertaken using multiple sample matrices and experimental conditions applying the automation and sample prep methods developed here. The profile of this assay as a rapid, easy, non-sensory, upstream quantification of freshness in versatile fish tissue matrices has broad utility, for example as a way to calculate shelf life metrics as shown here by a 3-month time course on raw, frozen shrimp.
Adenosine triphosphate (ATP) is a nucleoside composed of a ribose ring with an attached adenosine group and distinct hydrogen triphosphate bond. Among other critical biological functions ATP plays an important role in muscle contraction, and is an end product of cellular respiration. When cellular activity ceases ATP is naturally depleted through a combination of autolytic enzyme and bacterial action that reduce the hydrogen triphosphate and adenosine bonds, sequentially degrading ATP into smaller nucleoside and purine end products adensoine diphosphate (ADP), adenosine monophosphate (AMP), inosine monophosphate (IMP), Inosine (Ino), and hypoxanthine (Hx) respectively (Figure 1). Because these catabolites are produced early post-mortem their measurement in animal tissue can be used to determine quality control parameters such as muscle age and storage time, important indicators of freshness. The quantity of these nucleotides in tissue products can also influence sensory attributes, such as taste and texture.
Figure 1. Biochemical structure of freshness pathway characterized by ATP depletion to Hypoxanthine (Hx) following the arrest of cellular respiration post-mortem in muscle tissue.
Expressed as a K-value, freshness of fish tissue was first quantified by the Japanese team of Saito et al in 1959 using a 6 parameter equation that included all of the freshness pathway products. Lower K-values indicated fresher fish. This equation was later abbreviated by Karube et al in 1984 to a 3 parameter equation of Ki (%) for IMP, Ino, and Hx only.  %IMP and %Hx can be expressed as an equivalent freshness index by their relationship in Ki such that the higher the %IMP the fresher the fish. This relationship is illustrated by Figure 2, which also shows that the decomposition rate and concentration of these nucleotides over time is species dependent.
Figure 2. Highest levels of inosine monophosphate (IMP) are present during a window of time early post-mortem in muscle tissue, and decrease inversely to the increase of hypoxanthine (Hx). This profile can differ from species to species as shown for trout (left), mackerel (center), and sole (right).
In addition to sensory quality control methods that are based on subjective measurements by trained assessors, there are biochemical and chemical methods that quantitatively evaluate seafood freshness. These include immobilized enzymes on test strips, enzyme coated oxygen electrodes, coupled enzyme meters, colormetric and chromatographic analysis like HPLC among others. Some methods are incompatible with species of muscle tissue or process variables such as pre-cooking or canning, others are useful only in later stages of post-mortem (>= 10 or more days after death), or include costly equipment and more sophisticated technique employed by skilled personnel that are inharmonious with smaller onsite or field laboratories. It is generally recognized that the most relevant data comes from a combination of methods, or targets within a method. For example, if relative content of IMP were calculated at 50% from a sample extraction it would not indicate whether IMP was being generated or depleted, something clarified by knowing the percent of Ino and Hx that follow it.
To address some of these issues a novel enzymatic assay, the Precice® Freshness Assay, was introduced in January 2012 by French company NovoCIB (Figure 3).
Figure 3. NovoCIB Precice® Freshness assay principle and workflow. The assay works as 3 assays in one on up to 11 individual samples in a fixed 96- well microplate format with column 12 reserved for pre-filled standards (top left). Relative content of (%) IMP, (%) Hx, and (%) Ino is calculated following enzymatic conversion of nucleotides to NADH, and detection of NADH at 340 nm (lower left, and right).
Until the introduction of this assay there was no method available to measure IMP, Ino, or Hx in shrimp, lobster, or canned fish. Since the nucleotides are thermo stable testing can be done on virtually any species as unfrozen raw, previously frozen, frozen packaged, ‘on ice’, smoked, pre-cooked, prepared, or as a canned fish or other meat product, including shrimp and lobster. The mix and read format combines a rapid, objective quantified index of upstream freshness with an ease of use that allows in-house testing as an option to chromatographic methods. As a screening tool it can ascertain results for quality indicators before methods like Total Volatile Basic Amines (TVB), Trimethylainine (TMA) or Biogenic Amines (BA) become relevant. This versatility can be applied in testing up to 11 individual samples in duplicate on one 96-well microplate using the same sample prep method. This method uniformity allows the introduction of microplate automation that can offer improved assay workflow, throughput, and reproducibility over manual procedures (Figure 4).
Figure 4. BioTek Precision Power XS Liquid Handler (left) and Epoch Microplate Spectrophotometer (center) were used to validate automation and monochromator detection of the NovoCIB Precice® Freshness Assay. Epoch 2 (right) has also been verified to produce equivalent calculated results for the assay, and is fully compatible with the assay Gen5 sample file.
Originally designed using filter-based detection methods, here we validate monochromator detection of NADH at 340 nm using BioTek's Epoch™, including data comparability across BioTek's Epoch reader family (Epoch and Epoch 2). The low cost Epoch reader family offers a small bench top footprint that profiles nicely with smaller in house or field laboratories. Epoch also has the advantage of spectral scanning that allows visualization of NADH peak at 340 nm. Additionally, a kinetic read mode offers the option of detecting enzymatic conversion of the nucleotides to NADH over time. The Epoch 2 product line offers enhanced features such as linear and orbital shaking and a front loading vertical cuvette holder conducive to the Precice Freshness assay workflow options.
The security features of the BioTek Gen5™ software used to control the Epoch reader family make methods easily standardized and offer time-stamped audit trails of user and instrument activities, with some versions offering 21 CFR Part 11 compliance. The Epoch reader family can also be integrated with liquid handlers and robotics to increase the throughput capability for screening larger numbers of samples. As an example, potential workflow and throughput solutions of automating the liquid handling procedures of this assay using BioTek's Precision™ XS Microplate Sample Processor (PPXS) are demonstrated here. Instrumentation was validated over multiple runs using NADH standards and assay enzyme dilutions. Sample preparation options were then assessed for contribution to workflow efficiency and results optimization. Finally, a select method was applied to a time course study on frozen shrimp to observe the effects of storage time on nucleoside content at 6 points over a 3 month period.
Materials and Methods
- 1 L Pyrex Laboratory Beaker
- 50-mL screw cap conical tubes
- BD 10 mL Syringe Ref 309604
- BioTek 96-count pipette tips 200 μL (Catalog No. 98254; 98195)
- Brand ‘X’ imported farmed raw, frozen, peeled, whole jumbo shrimp (retail)
- Calibration solutions for Freshness Assay kit (NovoCIB PN K0700-001) containing B-Nicotinamide adenine dinucleotide (NADH) reduced form from ROTH Sochiel EURL [CAS 53-84- 9, C12H27,N7,O14,P2]
- Corning 150 mL Bottle Top Filter Polystyrene PN 431161
- Corning Costar Polystyrene 96-well round bottom microplate (PN 3797)
- Corning Polypropylene 96-well round bottom microplate (PN 3912)
- Millipore 50 mL Steriflip Cat No. SE1M003M00 • Pall Acrodisc Syringe Filter 0.2 μM 25 mm (PN 4192)
- Pall GHP Acrodisc GF, GF/0.45 μm, 25 mm (PN 4559)
- Potassium Hydroxide Sigma P-1767 (CAS 1310-58-3, KOH)
- Precice® Freshness Assay, NovoCIB PN K0700-003
- Sartorius Minisart Plus 1.2 μM Glass Filter + 0.2 μM membrane 28 mm syringe filter (PN 17823)
- Seahorse Polypropylene 12 Column Reservoir PN 201256
- Sterile single edge razor blades
- BioTek Epoch™ Microplate Spectrophotometer
- BioTek Precision™ XS Microplate Sample Processor (PPXS)
- BioTek Synergy™ H4 Hybrid Multi-Mode Microplate Reader
- BioTek Synergy™ HT Multi-Mode Microplate Reader
- BioTek Synergy™ Neo HTS Multi-Mode Microplate Reader
- BioTek Gen5™ Data Analsysis Software v2.01.14
- Centrifuge 50 mL tube compatible at 12,000 rpm or greater
- OMNI GLH, General Laboratory Homogenizer w/ Hard Tissue Omni Tip™ Plastic Generator Probes
- Vacuum Pump
- VWR Model 320 Heat Plate
1. Validation of Monochromator Detection and Automated Liquid Handling using NADH Standards
1 mL of 13 standards of NADH in 10 mM KOH were provided by NovoCIB at 0, 12, 27, 100, 125, 150, 204, 245, 305, 395, 480, 610, and 725 μM. An additional standard was prepared by first reconstituting 0.0129 gr KOH in 20.260 mL MilliQ water to obtain 10 mM KOH. This was used to dilute the 725 μM standard 13x to make a 55 μM standard. A standard curve was prepared by first dispensing 230 μL of each NADH standard in duplicate to 2 holding plates (Corning). An additional 8 replicates of NovoCIB and BioTek blanks (10 mM KOH) were added to the plate. One holding plate, a 96-count box of 200 μL pipette tips (BioTek, 98254), and a round bottom 96-well microplate (Costar) were loaded onto PPXS. The second holding plate was used for manual pipetting. For each standard two 100 μL dispenses were transferred from the holding plate to wells A04-H09 of the assay plate (Costar) in a 2 x 8 format for standards and 1 x 8 format for blanks using an 8-channel pipette, changing tips after each column. Both plates were read on four BioTek microplate readers in monochromator mode at 340 nm with the exception of the Epoch™ Microplate Spectrophotometer that also included a spectral scan from 300 to 400 nm in 10 nm increments. The NADH standard curve method was repeated twice over two days resulting in 2 runs of 4 plates as the volume of prepared standards allowed. Data was blanked on the average of the zero standards. The replicates of 10 mM KOH were used to calculate Limit of Blank (LoB). Standards were plotted using a linear curve fit, and precision, accuracy, Limit of Detection (LoD) and Limit of Quantification (LoQ) calculated. Accuracy was calculated by weighing each assay plate (Costar) pre- and post-dispensing then taking the delta weight.
Actual volume per well was determined by dividing the delta by the total number of dispensed wells on the plate and then multiplying by 1000. Percent accuracy was represented by the ratio of actual:expected volume multiplied by 100. Spectral analysis was graphed for the 0-150 μM standards. Standard curve data from Epoch was plotted in comparison to a competitor filter based microplate reader, competitor filter based cuvette spectrophotometer, other BioTek monochromators, and expected ODs as defined by the supplier of the NADH.
2. Automated Liquid Handling Optimization using Assay Reagents
Using kit extraction buffer as sample and diluent, 3 dilutions (1:1, 1:2, 1:10) of enzyme reagents were assayed over 3 runs for a total of 6 plates. For each run of 2 plates two operators in parallel prepared assay kit reagents according to the package insert resulting in 2.8 mL of each reagent (3 enzyme mixes and a reaction buffer), and 1 L of Extraction Buffer/kit. 1 mL of extraction buffer and 1 mL of each prepared reagent were added to four 50 mL conical tubes (Eppendorf) resulting in 2 mL of four 2-fold dilutions. 1.8 mL of extraction buffer was placed into another four 5 mL conical tubes followed by a 200 μL addition of each reagent resulting in 2 mL of four 1:10 dilutions. Remaining reagents were used to run 1x dilutions. Each operator dispensed 230 μL/well of each dilution to 4 wells on a holding plate (Corning) (8 wells total/run), then loaded 5 mL of extraction buffer to 3 lanes of a 12-channel reagent reservoir (Seahorse) (6 lanes total). 100 μL/well of extraction buffer was dispensed manually from the 12-channel reagent reservoir to columns 7-12 of an empty plate. Then the assay plates, holding plate, 12-channel reservoir, and 2 boxes of 96-count pipette tips were loaded to the Precision™ XS Microplate Sample Processor. Changing tips between columns, automation dispensed extraction buffer to cols 1-6 and reagent to cols 1-3 and 7-9 of each assay plate. The assay plates and holding plate were removed from the Precision XS Microplate Sample Processor and finished manually by dispensing 100 μL of remaining reagent from the holding plate to cols 4-6 and 10-12 of the assay plate. Plates were then read kinetically every minute for 1 hour at 340 using Epoch Microplate Spectrophotomer controlled by Gen5™ Data Analysis Software. Mean, standard deviation, and CV% were calculated for each dilution and each method within and between all 3 runs. Workflow was edited to allow continuous throughput for 1 up to ‘n’ plates in final assay mode. Workflow validation was achieved by calculating % error (100% - % accuracy) from a mock run of 6 plates using 200 μL/well of assay extraction buffer.
3. Sample Prep Method Optimization
Optimal sample prep workflow parameters were evaluated and validated on a single lot of Brand ‘X’ shrimp. 80 grams of sample were aseptically minced, pooled, weighed, placed in one of four 50 mL conical tubes, then diluted 1:2 (18gr:36mL) with kit extraction buffer. Two samples were homogenized for 45 seconds at speed 4 by slowing adding minced shrimp before capping all four samples, boiling for 20 mins, then cooling. Two other samples were centrifuged at 12,000 in 5 minute intervals, transferring exudate to a clean 50 mL conical tube after each interval until no pellet formed. Two other samples were filtered using either a vacuum driven bottle top filter (Corning) or 0.2 μm membrane syringe filter (Pall). 100 μL of each sample was dispensed to the assay plate in replicates of 4 for each nucleotide. The remainder of the assay was run according to the kit instructions. This was repeated on three additional samples without homogenization or centrifuging but clarifying using a 2-step vacuum filtration system (Millipore), a GF/0.45 μM 25 mm syringe filter (Pall), or a GF/0.2 μm membrane 28 mm syringe filter (Sartorius). Sample recovery volumes, materials, and number of steps for each clarification method were analyzed. A select method was then validated by first mincing, pooling, and weighing 25 grams of Brand ‘X’ shrimp at 45 days post purchase date, placing 4 gr into each of four 50 mL conical tubes, adding 8, 16, 32, and 40 mL (2x, 4x, 8x, 10x dilutions) of extraction buffer, capping the tubes, boiling for 20 minutes, cooling sample on ice for 5 minutes, recovering extract using syringe filtration (Sartorius), then assaying samples in replicates of 6 for each nucleotide per the kit insert. A pre-calibrated optimal dilution factor (2x) for shrimp (calculated by NovoCIB) was reproduced by running a 10 point serial volume dilution on a single 18 gram weight of minced, Brand ‘X’ shrimp at 68 days post purchase date diluted 1x in extraction buffer, boiled 20 minutes, then cooled on ice 5 minutes. 10 mL of 1x sample was filtered (BD, Sartorius) into a 15 mL conical tube as a starting standard. Two ¼ log dilutions were performed from 6 mL of start volume, followed by seven ½ log dilutions on each subsequent standard using extraction buffer. A zero standard was composed of filtered assay buffers only. Standards were transferred to the assay plate in duplicate for each nucleotide and NADH detected according to the kit insert (data not shown).
4. Applied Automation and Monochromator Detection
Optimized automated and sample prep workflow parameters developed during methods 2 and 3 were applied to testing Brand ‘X’ shrimp at 6 time points post purchase date (day 0, day 45, day 68, day 78, day 82, and day 88) using 2 lots of kit reagents (time 0 and 45 were tested with lot 1 reagents and days 68,78,82, and 88 with lot 2 reagents). At each time point 1 L of thawed extraction buffer was prepared according to the kit insert. 30 grams of shrimp were aseptically minced, pooled, and weighed into two 15 gram lots. Each lot was placed in a 50 mL conical tube, diluted 2x with extraction buffer (15gr:30mL), capped, boiled 20 minutes, then cooled on ice 5 minutes. While samples boiled, thawed assay reagents were prepared according to the kit insert. 2 mL of extraction buffer was loaded into lane 12 of a 12-channel reagent reservoir (Seahorse) for reconstitution of assay standards. 330 μL/well of reaction buffer (rows A,B), and enzyme mixes 1,2, and 3 (rows C-H respectively) were loaded into wells A01-H04 of a reagent holding plate (Corning). 30 mL of sample extract from each tube was pooled into a 100 mL beaker. 5 mL of pooled sample was aspirated into a 10 mL disposable syringe (BD). The syringe was fitted with a multi-layer disc filter (Sartorius) and sample was dispensed directly into lane 1 of the 12-channel reservoir. This was repeated 10 additional times for a total of 11 replicates. The reagent holding plate, sample reservoir vessel, an assay plate (NovoCIB), and 2 boxes of 96-count 200 μL tips were placed on the Precision XS Microplate Sample Processor supply deck. The Precision XS transferred 100 μL/well of each sample from lanes 1-12 of the sample reservoir to columns 1-12 of the assay plate respectively using the multichannel pipette, changing tips between each transfer. Again using the multi-channel pipette, the Precision XS transferred 100 μL/well from column 1 of the reagent holding plate to columns 1-3 of the assay, repeating for columns 2-4 of the holding plate to columns 4-12 of the assay plate respectively, changing tips after each transfer. The assay plate was removed from the Precision XS supply deck, shaken for 2 minutes, and read kinetically for 1 hour at 1 minute intervals at 340 nm on Epoch. Using Gen5 Data Analysis Software, the mean of each blank pair (see Figure 3 for the assay map) was subtracted from all wells in a column, followed by calculation of %IMP, %Hx, and %Ino. Using GraphPad Prism, a gaussian frequency distribution illustrating expected assay performance was plotted on blanks and standards from all experiments that were automated. LoB, LoD, and LoQ were calculated and compared to the same values obtained from the NADH standards run in Method 1. LoQ percent error was calculated as the sum of instrument bias (1% from 0-2.0 OD), imprecision (average precision (CV)), and total error (mean OD was used as the expected value) using raw ODs of blanks and NCs.
Results and Discussion
Figures 5 and 6 show results of comparative linearity of 0-610 and 0-725 μM NADH for monochromator and filterbased detection methods, and precision and accuracy of automated vs. manual pipetting of standards. Of note, intra- and inter-assay precision are all <2%, with an inter-assay precision delta of 0.452% favoring the automated method. Although the manual method had a higher average percent accuracy there was 12x greater variability in accuracy between plates than the automated method. Spectral scan data illustrates both the distinct peak of NADH detection at 340 nm and the tight correlation between manual and automated pipetting even at lower concentrations as shown for the 12-150 μM standards compared to the blank (0 μM). Although high end dynamic range differs for cuvette detection, linearity from 100 – 395 μM of NADH (0.5 <=OD <=1.8 OD) is correlative between detection methods with increased dynamic range of non-cuvette methods to >1.8 OD (>400 μM NADH) as represented by data for Epoch Microplate Spectrophotometer (also see Table 4 at end of document).
Figure 5. Linearity of 0-600 μM NADH compared between competitive detection methods (top), and for 0-725 μM NADH between monochromators of a same instrument family (bottom, BioTek Synergy line). Inset on left magnifies results for 0-55 μM, illustrating differences in method blanking. At right, linearity of NADH via monochromator detection had a linear range of 0.5 <=OD <= 2.5 (>50 - 610 μM) on all readers.
Figure 6. Precision and accuracy of automated vs. manual pipetting of 0-725 μM NADH standards using BioTek's Epoch™ monochromator detection (top). Even at low concentrations the distinct peak of NADH at 340 nm can be seen using Epoch spectral scan analysis (bottom, (A) utomated, (M)anual) as shown for 12-150 μM compared to 0 μM (each point n=4).
Table 1 shows inter-assay precision for automated liquid handling using assay reagent dilutions compared to manual. Although the variability for automated pipetting is higher by a delta of 1.548 CV%, manual pipetting resulted in 4x more outliers. Outliers on the automated method were due to insufficient reagent volumes resulting from the experimental design, and were expected as seen during reagent vessel loading. For the manual method this could have been compounded by either an observed higher reagent volume requirement for the 8-channel hand held pipette, leaving less than the required reagent volume for the last dispense, or operator inexperience preparing the reagent holding vessel volumes.
Table 1. Inter-assay precision for automated liquid handling validation as described by Method 2. Results reflect average from all reagents at each dilution for 6 plates.
A finalized continuous liquid handling throughput model was optimized for running the assay in real time, and Table 2 shows that reagent usage is sufficient across the microplate as shown by the elimination of outliers resulting from a mock throughput run of 6 plates.
Table 2. Accuracy results of automated workflow challenge test from mock 6-plate run. The PPXS specification for accuracy is <=2% error for each 50 - 100 μL dispense. The expected cumulative % error from two 100 μL dispenses is then <=4%.
Figures 7-9 and Table 3 show representative results from optimizing the clarification step of the assay sample prep workflow. Centrifuging and homogenizing sample resulted in high recovery of nucleotides, but only after repetitive cycles (Figure 7).
Figure 7. Comparison of results from different clarification methods during assay sample prep. Although NHC samples show the highest nucleotide recovery, they required 5 centrifuge cycles and exudate transfers before the absence of a tissue pellet. (F)iltered samples had to be batched into small volumes then repeatedly filtered using a 0.20 μM 25 mm syringe before enough clarified sample was pooled to run the assay.
Syringe filtration required the least number of steps and equipment, but single membrane syringe and vacuum methods at the membrane and disc sizes selected resulted in low to no recovery of sample depending on concentration (some data not shown). A filtration method consisting of a 10 mL syringe fitted with a multi-layer 28 mm disc (BD, Sartorius) was found to quickly recover the most sample over the broadest concentration range with the smallest pore size and required only one clarification step with no extraneous equipment (Table 3).
Table 3. Representative results for recovery of sample extract at different concentrations for comparative filtration methods.
This method was tested on 3 concentrations of sample (Figure 8), and an optimal sample prep workflow finalized (Figure 9).
Figure 8. Validation of final filtration method (10 mL syringe fitted with a GF+0.2 μM filter disc (BD, Sartorius)) using 3 dilutions of sample. Lower concentration of sample show a slight but expected decrease in available nucleotides, suggesting that the filtration method efficiently recovers sample even at higher concentration and that nucleoside detection is influenced more by dilution factor. 2x (4:8, above) is the optimized dilution factor for assaying shrimp as determined by NovoCIB.
Figure 9. Final optimized workflow. Freshness value equations are embedded into the Gen5™ sample file.
Figure 10 illustrates the freshness profile of Brand ‘X’ raw, frozen shrimp. The parabolic shift of Inosine (Ino) at this stage of product freshness reflects the decomposition dynamic as IMP is depleted in the tissue over time and Hx is formed.
Figure 10. Quantification of freshness pathway over 3 months for raw, whole, peeled, imported, farmed, frozen shrimp illustrating IMP decomposing to Ino that in turn decomposes to Hx. This product had a ‘best if used by’ date label that corresponded to 6 months (180 days) from the day of purchase.
Table 4 compares LoB, LoD, and LoQ, and linear and dynamic range for NADH standards in 10 mM KOH to those calculated from the assay reagents representing 13 and 9 separate automated runs of the assay over a 6 month period.
Table 4. Results for Limit of Blank (LoB), Limit of Detection (LoD), Limit of Quantification (LoQ), and linear and dynamic range for Epoch monochromator detection of NADH standards (0 – 725 μM) and assay reagents. LoB, LoD, and LoQ were calculated using methods in Clinical Laboratory Standards Institute protocol EP17-A. 
Figure 11 reflects the expected performance range of the assay using automated liquid handling and/or monochromator detection. Because positive controls are pre-loaded in the assay plate, each nucleotide has only 1 replicate of positive control. Blanks and negative standards could be run in replicates >1 when experimental conditions allowed. Lot 1 blanks and NCs are not included because of a plate change by the kit distributor. Because the counts were lower for the positive controls, the secondary y-axis was dimmed to maintain visual continuity. Spectral Scan capabilities of Epoch™ can be useful for discriminating NADH detection from possible assay reagent background for borderline samples (Figure 6).
Figure 11. Expected range of assay performance using automated liquid handling and monchromator detection. Raw blank values are plotted. Standard values (NC,PC) are blanked. Blanks and negative controls use the count on the left y-axis, whereas positive controls use the count on the mid y-axis (dimmed for visual continuity purposes).
Our validation data indicate that the NovoCIB Precice® Freshness Assay is amenable to BioTek automated liquid handling instrumentation and monochromator detection of NADH at 340 nm in 96-well microplate format offering standard operating procedure efficiency and diverse workflow, throughput, and detection options in most any laboratory environment.
 Goodrich, Wendy, Balakireva, Larissa, Challenge Testing of an Automated, Novel Freshness Assay Measuring Inosine Monophosphate (IMP), Inosine (Ino), and Hypoxanthine (Hx) as a % Ratio in Animal Muscle Tissue, BioTek Instruments, Inc. Application Note, Food Safety and Quality, 2014. http://www.biotek.com/ resources/app_notes.html
[2,3,4] Huss, Hans H. Quality and Quality Changes in Fresh Fish. FAO Corporate Document Repository, 1995. http://www.fao.org/docrep/v7180e/v7180e00.htm
 Tholen, Daniel W. M.S., CLSI EP17A Protocols for Determination of Limits of Detection and Limits of Quantitation; Approved Guideline, EP17-A standard published 10/20/2004 by Clinical and Laboratory Standards Institute. Second edition currently in print: Pierson-Perry, James F. et al. CLSI EP17-A2 Protocols for Determination of Limits of Detection and Limits of Quantitation; Approved Guideline - Second Edition, EP17A2E standard published 06/01/2012 by Clinical and Laboratory Standards Institute